We discuss the use of subdivision surfaces in the procedural design of imprint rolls for use in the roller imprinting process. Roller imprinting is being developed for the fabrication of microfluidic devices in polymer substrates. Imprint rolls are modeled using Catmull-Clark subdivision surfaces, and are procedurally designed based on feedback from finite-element simulations of the imprinting process. Microfluidic devices exhibit repeating patterns, and can be modeled using a small set of unique entities (or tiles). Imprint rolls are also modeled as a sum of tiles, and rolls are designed by studying the imprinting behavior of clusters of tiles corresponding to the repeating patterns seen in the device. This approach reduces the roll complexity and analysis time. The rolls need to be described in a sufficiently flexible format for the tile-based analysis to be effective. Conventional model representations are too cumbersome for piecewise iterative refinement as they require the manipulation of a large number of variables to modify surface features while preserving continuity. Subdivision surfaces, on the other hand, are naturally continuous and can be modified by manipulating a small number of variables. The ability to apply rule-based, arbitrary refinement on subdivision surfaces makes them especially suitable. The procedural modeling methodology and the subdivision design representation enable the integrated design, analysis, and manufacturing of imprint rolls, and has proven effective in decreasing the design-to-manufacture time of novel microfluidic technology.

For pt.I see ibid., p.735-45 (1990). A robot and positioning table system is a kinematically redundant system with respect to planar motion. Two strategies were developed in pt.I to resolve this redundancy and to specify path shapes that make the best utilization of the workspace and speed characteristics of the two devices. In this part, a one-variable dynamic programming approach is developed to obtain a near-minimum time and/or energy trajectory of the two devices. The developed strategies were studied using this path-planning algorithm on a model of a 3 d.o.f. robot and a two-axis linear positioning table for a variety of path shapes and constraints. In addition, experiments were performed in a typical workcell to show the feasibility of the developed strategies. In moving the two devices in opposite directions, the least travel time is obtained when the original path is resolved more in favor of the faster device, whereas the quality of resultant motion is dependent on the controller performance of each device.>

Coordinating the tool (robot) and workpiece (positioning table) motion in continuous manufacturing processes such as sealing, welding, and laser cutting offers several advantages over the motion of only the tool or workpiece. Better utilization of the speed and workspace characteristics of the two devices and improvement in tracking accuracy at sharp corners in the path are some of the merits of this approach. In coordinated motion, the two devices are simultaneously moved to track a given path, and two strategies are developed for coordinating their motion, where each is dependent on the given path information. The first strategy is applicable to any type of path and resolves the motion by splitting the displacement between any two points on the path into segments moved by the robot and the table. The robot and the table move in opposite directions in this strategy. The second strategy is applicable to sharp cornered paths and resolves the given path into two smooth paths. The first path is a double clothoid curve, whereas the second path is a tangential straight line path. The table executes a local motion around the corner in this strategy, whereas the robot is moved at all times. The uniqueness of this strategy in constructing the original corner path is proved. The proposed strategies are shown to be applicable to other coordinated motion systems.>

The design and characterization of a fast linear microactuator is described. The microactuator is constructed of number of piezoelectric plates that are placed on both sides of a sliding mass. Actuation of the plates in an accordionlike fashion provides a linear displacement of +or-0.36 mm. In addition, it has force capability of over 80 g, and is provided with both position and force-sensing capabilities. Characterization of the dynamic performance of the microactuator is presented, and initial tests on it show that it has a second-order system behavior with nonlinear response. The resolution of the microactuator is affected by the performance of the capacitive-type position sensor, and varies from 3 mu m to less than 0.6 mu m.>

The authors present a scheme that uses a feedforward neural network for the learning and synthesis task. Neural networks consist of a collection of interconnected processors that compute in parallel. The parallelism allows the network to examine various constraints simultaneously during the learning phase and enables reductions in computing time that are attractive in real-time applications. The learning abilities of these networks in a tuning operation are discussed. The network learns by observing the effect of the input variables of the operation (such as feed rate, depth of cut, and cutting speed) on the output variables (such as cutting force, power, temperature, and surface finish of the workpiece). The learning phase is followed by a synthesis phase during which the network predicts the input conditions to be used by the machine tool to maximize the metal removal rate subject to appropriate operating constraints.>